Loading...

Radar Target Recognition Using Range Profiles Synthesized by Random Stepped Frequency Radar

Sadeghi Ghartavol, Mohammad | 2020

484 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53400 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Bastani, Mohammad Hassan
  7. Abstract:
  8. Target recognition is one of the widespread applications of today's radars that requires obtaining target signatures using radar measurements. High-resolution range profile (HRRP) is one of these signatures that provides a one-dimensional radar image of the target. There are several methods for radar target's HRRP synthesis , all of which require a large bandwidth. One of these methods is the use of stepped frequency radar. One of the advantages of this method is providing a wide bandwidth by sending pulses with small bandwidths, thus obviating the need for broadband receivers and transmitters and making implementation easier. In many cases HRRP of target is sparse, because the number of strong scattering centers of target is significantly less than the number of range cells in HRRP, which allows the use of compressive sensing. Random stepped frequency radar is a type of stepped frequency radar that uses compressive sensing and in addition to reducing the number of pulses required to generate HRRP, increases the resistance to electronic attacks. One of the drawbacks of this type of radar is its high sensitivity to the target movement, which causes distortion in the reconstructed HRRP by adding some phases to the received echoes. Estimation and compensation of target motion parameters in this type of radar is one of the areas that need further research. In this thesis, a method based on maximizing contrast by searching in two-dimensional speed-acceleration space is proposed and it is shown that this method offers good accuracy in estimating the target velocity and acceleration. Also, the HRRP synthesized by this method is used to classify four targets using convolutional neural networks to compare its accuracy with stepped frequency radar and stationary target, which is considered as the ideal case
  9. Keywords:
  10. Speed Estimation ; Convolutional Neural Network ; Compressive Sensing ; Range Profile Function ; Radar Target Recognition ; Target Acceleration Estimation ; Target Velocity ; Random Stepped Frequency Radar

 Digital Object List

 Bookmark

No TOC